In 2015, a study looked at drug-industry sponsored clinical trials versus those funded by the NIH, and concluded that about 20% of industry trials did not report results when required to do so, while the figure for NIH-sponsored research was about 50% (and even more for trials funded by other institutions). At the time, I faulted a lot of the reporting on this issue, since an eye-catching figure of only 13% of pharma trials reporting kept coming up (without being corrected for the trials that were under no legal obligation to report at all during the period studied), and I particularly noted Ben Goldacre as someone who had failed to make that distinction.
Now we have similar figures for Europe, thanks to. . .Ben Goldacre, among others (commentary here at Nature). Regulations in the EU for reporting data within a year of trial completion have only been put in place recently as compared to the US (although, to be fair, they seem more straightforward in the EU), but this is a look at how that’s going. And it turns out that (of the trials that were required to report results) 32% of industry-sponsored trials have missed the deadline, while 89% of academic-run trials have failed to report. That latter figure is particularly bad, but it’s worth noting that both figures are worse than the US ones from 2015 (which may well have improved by this point). I’m glad to see that Goldacre is using numbers that are corrected for obliged-to-report trials this time, anyway. This paper does go into those details, which is good (but so did the 2015 one, and that didn’t stop its results from being misreported).
The take-home from both studies is that most clinical trials (in both the US and EU) are now obliged to report data within a year of finishing. Many trial sponsors seem to be having trouble making that deadline, and so far the problem seems to be worse in the EU. And on both sides of the Atlantic, it’s the drug-industry trials that have a far better compliance rate compared to the academic ones.
Here’s another interesting way of breaking down these numbers: if you look at the major sponsors (those with 50 or more trials to their names in the database), only 11 of them have a 100% reporting rate. All of those are from industry. Meanwhile, 32 of those big players have a flat zero per cent reporting rate, complete noncompliance. All of those are academic institutions. This is not a good look. The paper also notes “extensive omissions and contradictory data” in the EU registry entries, on top of the overall reporting issues. The hope is that since these requirements are relatively new that things will improve, but on the other hand, that will only happen if someone points it out and asks for better. Or, if that doesn’t work, if someone hammers on people a bit when they don’t improve. We’ll see if it gets to that point.